diff options
Diffstat (limited to 'SA_POP/ChangeLog')
-rw-r--r-- | SA_POP/ChangeLog | 41 |
1 files changed, 39 insertions, 2 deletions
diff --git a/SA_POP/ChangeLog b/SA_POP/ChangeLog index 329afa7fbf1..6e86624ada9 100644 --- a/SA_POP/ChangeLog +++ b/SA_POP/ChangeLog @@ -1,3 +1,36 @@ +Tue Jan 19 19:01:00 UTC 2010 John S. Kinnebrew <john.s.kinnebrew@vanderbilt.edu> + + * Planner.cpp: + + Clean-up on Planner::calc_plan_eu() + + * experiments/EU_Performance/Exp_Core/Exp_EU_Planner.h: + * experiments/EU_Performance/Exp_Core/Exp_EU_Planner.cpp: + * experiments/EU_Performance/SAPOP_Exp_EU.cpp: + + Clean-up on EU experiment code, including throwing out plans with + invalid expected utilities (due to inability of SA to catch certain + double loops). + + +Tue Jan 19 07:03:24 UTC 2010 Ben Podgursky <bpodgursky@gmail.com> + + * Planner.cpp: + * experiments/EU_Performance/Exp_Core/Exp_EU_Planner.cpp: + * experiments/EU_Performance/SAPOP_Exp_EU.cpp: + + Additional debug output (checking for > 1 probabilities) + +Tue Jan 19 02:07:16 UTC 2010 Ben Podgursky <bpodgursky@gmail.com> + + * experiments/EU_Performance/Exp_Core/Exp_EU_Planner.cpp: + * experiments/EU_Performance/SAPOP_Exp_EU.cpp: + + Turned off memory leak checking (took ages to output the leaks after long runs) + + Check for invalid plan EU values. + + Sat Jan 16 22:46:36 UTC 2010 John S. Kinnebrew <john.s.kinnebrew@vanderbilt.edu> * SA_POP_Types.h: @@ -5,8 +38,12 @@ Sat Jan 16 22:46:36 UTC 2010 John S. Kinnebrew <john.s.kinnebrew@vanderbilt.ed * experiments/EU_Performance/Exp_Core/Exp_EU_Planner.cpp: * experiments/EU_Performance/SAPOP_Exp_EU.cpp: - Experiment ignores (just SA, no planning) initial conditions/goal that result in invalid goal condition probabilities (or ones below the #define’d threshold). - Experiment ignores generated plans with the same EU (to within a #define’d threshold percentage) as the initial plan. + Experiment ignores (just SA, no planning) initial conditions/goal that + result in invalid goal condition probabilities (and ones below the + #define’d threshold). + + Experiment ignores generated plans with the same EU (to within a + #define’d threshold percentage) as the initial plan. Sat Jan 16 19:57:13 UTC 2010 John S. Kinnebrew <john.s.kinnebrew@vanderbilt.edu> |